Court blocks Perplexity's Comet browser from Amazon's accounts

A federal court on March 9, 2026 granted Amazon a preliminary injunction blocking Perplexity’s Comet AI agents from accessing Amazon password‑protected accounts (Case No. 25-cv-09514-MMC); the order is stayed seven days for an appellate stay and requires Perplexity to destroy any Amazon customer data obtained and certify compliance. The court found Amazon likely to succeed on CFAA (18 U.S.C. §1030(a)(2)) and California Penal Code §502(c)(7) claims, tied damages to remediation costs (court found Amazon incurred more than $5,000), and denied Perplexity’s request for a $1bn bond and a stay. The ruling materially constrains Comet’s agentic commerce value proposition (Comet launched July 9, 2025; initially $200/month Max plan) and reinforces platform control over AI agents — relevant for ad monetization given Amazon Ads generated $17.7bn in Q3 2025 (+22% YoY).

Analysis

This ruling creates a legal moat around platform-controlled identity and transaction surfaces that materially increases take rates for platform owners over the medium term (3–24 months) by raising the cost of third‑party agentic commerce. Expect enterprise and retail platforms to require agent identification headers, contractual API keys, or server‑to‑server purchase rails; each costs agent builders time and engineering dollars and favors incumbents that can monetize compliant agent integrations.

Near-term (days–weeks) the key binary catalyst is the Ninth Circuit stay decision — denial crystallises lost TAM for agent-first browsers in the largest e‑commerce channel and forces data destruction/technical remediation costs onto Perplexity and peers. Over months, the precedent accelerates vendor demand for bot‑management, device‑auth, and consent audit trails, creating a revenue opportunity for security/CDN vendors that can sell turnkey agent governance (expected contract cycles 3–9 months).

Countervailing paths exist: a settlement or technical remediation (unique agent UA, tokenized OAuth flows, client‑side credential vaulting) would restore much of agent utility within 3–9 months and blunt the injunction’s commercial permanence. Longer term (1–3 years), regulation or industry standards (IAB Agentic RTB adoption, platform APIs) could institutionalise agent access rules, splitting market winners between platforms that capture ad/commerce margins and middleware providers that standardise secure agent integrations.

Key Decisions for Investors

• Long AMZN (stock or 3–6 month call spread): Buy AMZN outright or buy a 3–6 month call spread to capture an expected re‑rating as Amazon monetizes tighter agent controls and reduces ad fraud/noise. Time entry after the Ninth Circuit denies a stay (or within 48–96 hours if denial appears likely). Risk: regulatory reversal or rapid agent compliance could pare upside; reward: 6–12% upside in 3–6 months if ad yield lift persists.
• Long NET (Cloudflare) 6–12 month horizon: Buy NET to play higher demand for bot management, edge auth, and logging as platforms and retailers accelerate contracts. Target 15–30% upside if incremental ARR ramps from managed bot solutions; risk is slower enterprise procurement cycles and margin pressure from competition.
• Tactical long RDDT (3–9 months) size‑small: Buy RDDT to reflect strengthened IP/control rights for platforms following this precedent, which supports monetization and enforcement against large scrapers. Keep position small (2–3% portfolio) given legal variability; reward is re‑rating if platform enforcement increases subscription/advertising quality.